Abstract If the main effect of long-term exposure of tomato plants to salinity is the accumulation of toxic concentrations of Na + and Cl − in the leaves, then the selection of ‘excluder’ rootstocks should increase tolerance to salinity in grafted tomato plants, independently of the genotype used as the scion. The question addressed in this study is whether shoot genotypes with an ‘excluder’ character are able to increase their salt tolerance when grafted onto rootstocks of the same characteristics. Moneymaker (with excluder character) was grafted onto two root genotypes, Radja and Pera, selected for their very different ability to regulate the transport of saline ions to the shoot over time. Grafting onto either Pera or Radja improved fruit yield compared to the self-grafted plants of Moneymaker (M/M) when the plants were grown at 50 mM NaCl, whereas there was no effect of either rootstock or of grafting per se (M/M) on fruit yield in the absence of or at 25 mM NaCl. The relationship between the salt responses to mid- and long-term depended on the stress level; after 27 d of 150 mM NaCl treatment, both graft combinations enhanced similarly their salt tolerances as did in the long-term experiment. Moreover, the tolerance induced by rootstock was related to the low rates of saline ion accumulation in their leaves. However, the positive effect of rootstock was only observed with rootstock Pera when the grafted plants were grown at 50 mM NaCl (the same salt level used in the long-term experiment) for 35 d. According to the physiological changes induced by rootstock in the leaves, the different salt responses seem to be due to the fact that the osmotic effect predominated on the toxic effect under these last conditions. Consequently, in order to select rootstocks care must be taken in the timing of any selection process: the stress level and length of exposure to salinity must be sufficient for the true differences in salt tolerance for toxicity to be shown. Taken together, these results show the effectiveness of grafting to enhance fruit yield in tomato and provide evidence that the positive effect induced by rootstock is related to the re-establishment of ionic homeostasis.